This year’s economic growth in Sweden has been impressively strong. GDP growth in the first half-year, on the first half of 1998, was 3.4 per cent. The number in employment has risen by approximately 100,000 persons. As many people are returning to the labour market now that more jobs are becoming available, the reduction of unemployment has not been as marked. But total as well as registered unemployment are falling from the previously high levels.
This has been accompanied by price increases that are moderate. To date this year, inflation’s underlying rate—measured by UND1X, which excludes interest expenditure, indirect taxes and subsidies—has fluctuated between 1.0 and 1.5 per cent.
The picture since the summer seems to point to a continuation of favourable growth and employment, with no immediate threats to price stability. But we have to be on our guard, particularly as the economy is now in an upward phase. Due to the time lag in monetary policy, the Riksbank’s interest rate decisions have to be based on the prospects fairly far ahead. The most relevant period for our deliberations is the outlook for inflation twelve to twenty-four months ahead.
Looking back
The current picture of the Swedish economy is very different from the situation at the time of this assembly last year. The mood was then clouded by the Asian crisis, Russia’s suspension of debt payments, the problems with the hedge fund Long Term Capital Management and an international share price fall.
The hedge fund problems aroused fears that banks in general would become so cautious about providing credit that this by itself would paralyse the global economy. That such a credit crunch was in fact imminent last autumn was indicated by the widening interest rate spreads to borrowers with good and less good credit ratings.
Last autumn’s financial crisis accordingly added to the negative real economic effects that were expected to come from the crisis in Asia. Forecasters reassessed the world economic outlook and revised their growth and inflation forecasts downwards. Central banks around the world did the same and the Riksbank was no exception. Instrumental rates were cut in many countries. Between June 1998 and spring 1999 the Riksbank reduced the repo rate in a number of steps from 4.40 to 2.90 per cent, that is, by a total of 150 basis points.
The repercussions of the financial disturbances that have hit the world economy in recent years do seem to have been moderated by the actions taken by central banks. To some extent, the process also tends to rectify itself. So all in all, the economic prospects for the coming years have brightened again, for the world in general as well as for Sweden.
But this does not mean that the international economy has been inoculated against new, unexpected shocks. Remember that no one predicted the extent and course of the Asian crisis. Neither were the risks that led to last autumn’s international financial turbulence highlighted in advance.
Our knowledge of phenomena of this kind is limited. It may be asked whether it will ever be possible to predict such matters accurately. The models with leading indicators that are being tested at present tend to be too blunt. They are liable to sound the alarm about things that never occur and miss those that do happen.
The picture is bright at present but we know that some countries are in a vulnerable position that can suddenly deteriorate. Stock market prices are notably high, not least in the United States. The existing imbalances in the US economy may have effects in the future. But as the course of such events is genuinely difficult to foresee, central banks have little option but to deal with a new situation if and when it arises.
Repo rate too low in a longer perspective
After the series of reductions, the repo rate is currently 2.90 per cent. Over a complete business cycle, such a level would probably be too low, even in an economy where inflation expectations have settled down around the targeted rate of 2 per cent.
So the present monetary stance can be said to be relatively expansionary. With the budgetary measures that have been proposed, the fiscal stance will also exert an effect on demand that is more expansionary in the future. In the absence of negative shocks, such a direction of policy normally tends to generate GDP growth above the rate that the economy is capable of sustaining in the longer run. If such a situation is allowed to continue for too long, it can lead to problems with bottlenecks and inflationary tendencies. In time, expectations of inflation’s level in the longer run may be adjusted upwards and this can result in a price and wage spiral that brings inflation above the targeted rate.
Such a course of events has to be prevented by the Riksbank, which the Riksdag (Sweden’s parliament) has made accountable for maintaining price stability. So sooner or later the level of the repo rate will have to be normalised in order to counter an acceleration of inflation in an upward cyclical phase. This is something that the Riksbank is discussing, as was evident from the published minute of the Executive Board meeting on 12 August. In the present circumstances what we have to do is decide when—not whether—the time has come to initiate the shift towards a less expansionary monetary stance.
As regards the timing of this type of monetary policy measure, it is generally true to say that if the Riksbank waits too long before making a necessary repo rate adjustment, when the increase does come it may have to be larger than otherwise in order to arrest the overheating and inflationary tendencies. That could introduce undue instability into output and employment.
A higher repo rate does not necessarily impede the continuation of a favourable economic trend; it may even be required for a good and sustained development of employment and growth, as is illustrated by how the Federal Reserve has acted in the United States. A judicious series of increases in the federal funds rate seems to have contributed to the longest post-war period of growth in the United States. Since it was raised by degrees from 3 per cent from 1994 to 1995, the American instrumental rate has moved between 4.5 and 6 per cent.
Having said this, it must be realised that there is no simple answer to the question of what constitutes a judicious timing or magnitude of a repo rate increase. One alternative, of course, is to act early and make a small increase even though indications of growing inflationary pressure have not actually shown up. Such a line can be seen as insuring against the adverse consequences an acceleration of inflation might entail. In my opinion, however, there are other aspects that it is important to analyse and incorporate in the assessment.
Lower inflation propensity?
One reason for greater circumspection today in monetary policy is that the trade-off between growth and inflation may have changed. During the 1990s the paths of inflation and real economic activity in many countries, Sweden included, have not followed the pattern that their earlier relationship suggests.
This has started a discussion, in the United States and elsewhere, of whether macroeconomic structures have changed so that higher growth is now feasible without inflationary tendencies. It is sometimes argued, for example, that one of the forces behind today’s improved economic performance in the United States is the rapid advances in computer technology and telecommunications; this has led to more efficient processes for production and distribution and thereby paved the way for higher productivity growth.
Productivity in the United States does appear to have been rising more strongly. Some observers consider that the potential annual growth rate may have moved up from 2 to up to 3 per cent, perhaps even more.
In Sweden, too, there are many indications that the growth of labour productivity in the 1990s has been higher than in the 1980s. The upward shift may perhaps amount to half of a percentage point. It is this that has prompted the common, but not entirely certain, estimate that the Swedish economy can now sustain an annual growth rate between 2 and 2.5 per cent. But Swedish statistics do not yet confirm that productivity and thereby potential growth have increased fairly appreciably, as has been the case in the United States. In time, however, the pace of developments in computerisation and telecommunications will probably also have a major impact on production and distribution systems in Sweden. Sweden is well to the fore in the use of the Internet. Still, it is not yet clear when these developments will exert a more general effect on growth.
A point worth noting is that the upward shift in US productivity growth occurred fairly late in the business cycle. The Swedish economy has not reached that stage. Over a normal cycle, without any impulses from an improvement in potential growth, statistics show that productivity gains tend to slacken towards the end of the upward phase. That was the case in Sweden, for example, in the late 1980s.
Now that low inflation has been established, it is conceivable that the Swedish economy functions differently, so that productivity growth might also be better here in the coming years. The example of the United Kingdom, however, suggests otherwise. Activity there has been relatively stronger than in most other European countries but clear signs of better productivity growth are still not discernible.
If there is no direct evidence that it is appreciably higher productivity that lies behind the decreased inflation propensity in Sweden in the 1990's, where should we look for an explanation? I want to discuss two possibilities: the inflation process as such and the level of unutilised resources. At the same time I shall consider their significance for future inflation.
Has the inflation process changed?
Perhaps the most natural explanation for the fact that inflation in the 1990s has been lower than its historical relationship with growth suggests, seems to be the markedly lower inflation expectations. This is also indicated by data from studies at the Riksbank.
From a level between 6 and 8 per cent in the 1970s and ‘80s, in recent years expectations of longer-term inflation have gradually moved down to about 2 per cent. This is an impressive change that stems in the main from regime shifts in various fields that have been achieved since the late 1980s. Furthermore, it looks as though expectations are now anchored relatively firmly in the vicinity of that level.
Once a low-inflation regime has been established and people count on price increases being low in the longer run, the rate of inflation is determined mainly by the demand situation in relation to production capacity. During a process of disinflation, however, low capacity utilisation and falling inflation expectations both exert a downward effect on actual inflation. This can give the impression that the relationship between demand and inflation has changed. In that the pattern in the 1990s has been dominated by the reduction in expected inflation, the partial relationship between demand and inflation has been difficult to analyse, not least because it is hard to arrive at statistical measurements of the output gap, structural unemployment and their impact on the inflation process.
Still, the hypothesis that inflation’s sensitivity to resource utilisation has changed from the 1980s to the ‘90s does seem to have some support. Anecdotal information as well as more structured studies point in that direction. Perhaps it is the case that the inflation process is being influenced by the low-inflation environment that now prevails in Sweden. Many studies have found that low inflation is associated with a low variability in the rate of inflation. What we do not yet know is whether this lower sensitivity to demand variations is completely symmetric. It could, of course, be bound up with low resource utilisation, so that the situation is different when GDP is closer to the potential growth rate.
What is the level of unutilised resources?
While the downward shift in inflation expectations seems to have predominated, another factor behind the reduction of inflation has been the low level of resource utilisation after the crisis in the early 1990s. Demand growth is now above the rate that is sustainable in the longer run. This naturally brings the level of unutilised resources into the centre of the inflation analysis.
In an economy that is expanding rapidly, problems with bottlenecks may occur temporarily because new production facilities are not always available just when they are needed to cope with increased demand. However, experience in the 1990s suggests that firms are more flexible than before, both in the utilisation of existing capital stocks and in the procurement of new technology, machinery and plant when demand rises. So, if anything, it is not the level of unutilised capital resources that is critical at present.
Instead it is probably the labour market that should be focused when analysing resource utilisation. Another reason for doing so is that in a price analysis at macro level, the weightiest item in costs is wages: total wage costs are equivalent to over two-thirds of GDP.
The level of unemployment, although it has been reduced, is still high, particularly compared with the post-war period up to the early 1990s. Seen from that angle, some might say that the Swedish economy has plenty of unutilised resources.
Experience from other European countries has taught us, however, that when an economy has been through a severe shock, as happened in the early 1990s, unemployment tends to stick at a high level, at least for a while. This structural or equilibrium unemployment cannot be reduced by means of expansionary economic policies. Attempts to do so simply lead in the end to a higher inflation trend and probably an economy that functions less efficiently. This is because wage costs tend to adjust upwards as the expansionary policy acts on the economy and higher long-term inflation expectations are established.
Viewed in this way, unemployment could be divided into two components, structural and cyclical. If the level of structural unemployment could be identified with some precision, it would be easier to gauge how much room the Swedish economy still has for expansion above its potential growth trend.
As always when one turns from theory to practical policy, things immediately become much more difficult. You have to make do with estimates, which in the nature of things are uncertain and therefore have to be interpreted with caution.
Still, the available estimates do suggest that in Sweden the equilibrium level of unemployment rose in the early 1990s and is somewhere in the interval 4.5 to 7.5 per cent, expressed in terms of registered unemployment. The OECD also concludes that the level has increased markedly in the 1990s and judges that in Sweden equilibrium unemployment is around 6 per cent.
Not everyone agrees, however, that equilibrium unemployment is markedly higher than before. But even though we cannot be certain about its exact level, an increase is indicated by the fact that otherwise the currently high unemployment would be acting as a strong restraint in wage formation. Instead, the wage level in Sweden is still rising somewhat faster than in our main competitor countries.
Considering that the rate of registered unemployment is currently around 6 per cent and thus inside the estimated interval for structural or equilibrium unemployment, this might suggest that the Swedish economy is relatively close to full capacity utilisation. It is important, however, that the Riksbank’s monetary policy does not unnecessarily hamper decreased unemployment, particularly as the estimates of equilibrium unemployment are somewhat uncertain. So all the other information about labour market tendencies should also be considered.
When the economy is moving closer to its capacity ceiling, certain statistical phenomena can be expected to occur more clearly and frequently. The same applies to anecdotal information in such a situation. What I have in mind is, for example, bottleneck problems in the labour market and signs of difficulties in recruiting skilled labour. Another phenomenon is an increased propensity to resort to labour disputes, in the form of regular strikes or wildcat behaviour. Moreover, firms may have to compromise over job qualifications and take on less skilled personnel in order to cope with demand for their goods or services.
Some such signs have been observed in recent months. Firms are already finding it difficult to recruit computer consultants, for instance, as well as some other categories of personnel. The situation in some parts of Sweden is also more strained than in other parts. This suggests that the economy is moving towards a situation where firms find it increasingly hard to keep up with demand for goods and services; a situation, in other words, where demand is rising too rapidly. At the same time, the bottleneck problems are still fairly limited. There are no signs as yet of more widespread shortages that might generate unsettling inflationary impulses. The matching of job applicants and vacancies is still proceeding relatively smoothly and the wage statistics to date this year show increases that are a little lower, if anything, than expected. This may imply some respite for monetary policy.
Conclusion
Tomorrow the Riksbank’s Executive Board will be meeting to make three decisions. The first concerns the adoption of this year’s third Inflation Report. Then we are to agree on what I shall say to the Parliamentary Standing Committee on Finance on Wednesday when I present the Board’s appraisal of monetary policy. Last but not least we will discuss the significance of the inflation analysis for the future repo rate.
Our third decision is whether to raise the repo rate or still leave it unchanged for the time being. Today I have discussed some issues that I am sure will also be raised at the Board’s meeting tomorrow. It will be a matter, as always, of composing an overall picture out of all the indicators of inflationary pressure that the Riksbank studies. With the new Executive Board, this also involves combining how six individuals view this picture so that we arrive at a joint monetary policy decision.
But I will not anticipate the outcome of those discussions and compositions—that is something that will have to
